Hi, The magnetic field at HF restricts the current paths to the outer contours. This has two effects. One as you mentioned is the longer path. This has a two fold effect it gives dispersion and insertion loss a very degrading effect for data. Other is that the contour will have many impurities and sharp ends whic will further increase the loss. This is corroberated by measurement. As a microwave engineer in the past it was necessary to to make fine deposits on the surface of tracks and waveguides to reduce these losses. This gave a mirror finish. The highest frequency I worked at was 250 GHz where it was ever more evident. Hope this can give you some understanding of the order of things. Jim "chen, jinhua" wrote: > Jim > > Could you share the data with me? I did some test recently. The > skin effect loss is not as bad as I expect at 10 Gbps data rate. > > I have two pairs of differential traces with 100 ohms differential > impedance. One pair is microstrip line with 8 mils line width. The > second pair is stripline with 5 mils line width. The trace lengths > are 6" and 12" long for each pair. The eye patterns were took with > 10 Gbps PRBS data pattern. The difference of eye patterns between > two pairs is so small. It let me to believe that the skin effect > loss is not the significant part at 10 Gbps. > > The roughness of copper sufface is two dimentional. One dimention > increase the length, which increases the resistance. And the other > dimention increase the perimeter of copper cross section, which > decreses the resistance. So, they should cancel each other. This > is just my 2 cents. > > Regards! > > Jinhua Chen > Consulting Design Engineer > 176 South St. > Hopkinton, MA 01748 > 508-249-4527 > Pager: 8779700234 > chen_jinhua@xxxxxxx > > EMC² > The Enterprise Storage Company > > -----Original Message----- > From: Jim G Roberts [mailto:jgroberts@xxxxxxxxxx] > Sent: Wednesday, February 12, 2003 9:23 AM > To: milabont@xxxxxxxxx > Cc: 'Si-List' (E-mail) > Subject: [SI-LIST] Re: dielectric loss question > > Mike, Peter, > It is also interesting that since data rate are in the 10GHz rates > that > the > surface roughness begins to give a large percentage of the losses. > These are also very frequency sensitive. > If of interest I have a PPT file. > > -- > Regards, __________ James G Roberts > /___ ____ | jrobert@xxxxxxxxxxxxxxxxxxxxx > Jim __ / /___/ / jgroberts@xxxxxxxxxx > / /_/ /---| | Room: BE436, Hilversum > \____/ /_/ Tel: +31 35 687 4308 Fax: 5976 > > Mike LaBonte wrote: > > > I don't have any info specific to moisture in PCBs. 2.45GHz > > is where the dielectric permittivity of water begins to decrease > > significantly, but it is not a peak absorption frequency. Peak > > dielectric loss for water occurs in the range of 10GHz to 160GHz, > > depending on the temperature. The peak frequency is higher for > > higher temperatures. At 25degC the peak is near 40GHz. > > > > At DesignCon 2000 Howard Johnson presented "Multi-Level Signaling", > > in which he gave the rate of change for signal degradation due to > > skin effect and dielectric loss at 20dB/decade near 1-2GHz. At > > least half of that is dielectric loss. But the paper also shows > > noise increasing at 20dB/decade, and SNR falling to zero under 10GHz. > > > > My take on it is that water will not be the main cause of signal > > loss. > > > > http://www.sbu.ac.uk/water/microwave.html > > http://signalintegrity.com/Pubs/misc/mls.htm > > > > Mike LaBonte > > > > -----Original Message----- > > From: si-list-bounce@xxxxxxxxxxxxx > > [mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of Peter Arnold > > Sent: Tuesday, February 11, 2003 2:35 PM > > To: 'Si-List' (E-mail) > > Subject: [SI-LIST] dielectric loss question > > > > All, > > > > A question from the curious - in high-speed signaling environments we may > > create time-varying electric fields that have components close to the > > frequency used to heat water in microwave ovens, somewhere around 2.45GHz. > > This suggests a portion of dielectric loss at these frequencies might be > due > > to dielectric heating of any moisture that might be present in the > > fiberglass. > > > > Is this in fact the case, and what magnitude is the effect at normal > > humidities? Does loss increase in real multi-GHZ signaling systems as > > dielectric moisture content increases? > > > > Thank you, > > peter arnold. > > > > ------------------------------------------------------------------ > > To unsubscribe from si-list: > > si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field > > > > or to administer your membership from a web page, go to: > > //www.freelists.org/webpage/si-list > > > > For help: > > si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field > > > > List archives are viewable at: > > //www.freelists.org/archives/si-list > > or at our remote archives: > > http://groups.yahoo.com/group/si-list/messages > > Old (prior to June 6, 2001) list archives are viewable at: > > http://www.qsl.net/wb6tpu > > > > ------------------------------------------------------------------ > > To unsubscribe from si-list: > > si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field > > > > or to administer your membership from a web page, go to: > > //www.freelists.org/webpage/si-list > > > > For help: > > si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field > > > > List archives are viewable at: > > //www.freelists.org/archives/si-list > > or at our remote archives: > > http://groups.yahoo.com/group/si-list/messages > > Old (prior to June 6, 2001) list archives are viewable at: > > http://www.qsl.net/wb6tpu > > > > ------------------------------------------------------------------ > To unsubscribe from si-list: > si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field > > or to administer your membership from a web page, go to: > //www.freelists.org/webpage/si-list > > For help: > si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field > > List archives are viewable at: > //www.freelists.org/archives/si-list > or at our remote archives: > http://groups.yahoo.com/group/si-list/messages > Old (prior to June 6, 2001) list archives are viewable at: > http://www.qsl.net/wb6tpu > -- Regards, __________ James G Roberts /___ ____ | jrobert@xxxxxxxxxxxxxxxxxxxxx Jim __ / /___/ / jgroberts@xxxxxxxxxx / /_/ /---| | Room: BE436, Hilversum \____/ /_/ Tel: +31 35 687 4308 Fax: 5976 -- Binary/unsupported file stripped by Ecartis -- -- Type: application/pdf -- File: Surface_roughness.pdf ------------------------------------------------------------------ To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field or to administer your membership from a web page, go to: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu